SpaceX's Secret AI Division Just Hit $7.6 Billion — And Now Elon Musk Wants Tesla

Kind: captions Language: en Three months ago, the most famous rocket company on Earth made more money from computers than from rockets. Last quarter, it went much further. On October 20th, 2026, SpaceX reported that its artificial intelligence business brought in 7.56 billion in a single quarter. A year earlier, that same business made about $813 million. That's an increase of 830% in 12 months. For the first time since SpaceX went public, the AI side of the company made more money than Starlink, the satellite internet service that has been SpaceX's financial engine for years. It also made more than four times what the company earned from launching rockets. Then on the same evening, an analyst on the earnings call asked Elon Musk the question everyone had been waiting for. Is SpaceX going to merge with Tesla? Musk said he felt like he was tap dancing through a minefield. Then he said something he had never said so plainly in public. Both SpaceX and Tesla have formed special committees of independent directors to discuss combining the two companies. And in his words, he thinks it would be a good idea. If it happens, it would combine a company the market values at about $2 trillion with one worth about 1.5 trillion. That would create one of the most valuable companies on the planet with one man at the center of all of it. This is the story of how a rocket company quietly turned into one of the biggest AI infrastructure businesses in the world in less than a year. It's not a secret since every number I'm about to give you is in public filings, but most people are still picturing SpaceX as a rocket company. I'll show you where that 7.6 billion actually comes from. And the answer surprises a lot of people because most of it doesn't come from Grock. I'll show you the staggering amount of money SpaceX is spending to keep it going. And I'll walk you through exactly why Musk wants Tesla, what stands in the way, and the one vote that could decide whether it happens at all. If you like coverage that sticks to what's actually confirmed and clearly tells you what isn't, subscribe. This channel runs on real filings, real numbers, and no invented drama. A quick note before we start. This is not financial advice. I'm not telling you to buy or sell anything. I'm I'm telling you what happened and why it matters. To understand how we got here, you have to go back to the start of this year when SpaceX was still a private company and XAI was a separate startup that was losing money at an alarming rate. Musk founded XAI in 2023. It built the Grock chatbot and in 2025 it absorbed the social media platform X through a share swap. It was growing fast, but it was burning cash even faster. SpaceX's own IPO filing later showed just how deep the hole was. In 2025, XAI lost $6.36 billion on revenue of $3.2 billion. It spent 12.7 billion on capital expenditures that year, mostly on the chips and data centers needed to train AI models. That was more than SpaceX spent on Starlink in its rocket business combined. In the first quarter of 2026, XAI lost about $2.5 billion from operations on just $818 million of revenue. One analyst at PitchBook put it bluntly. If you compare XAI to a traditional software company, he said the financials look reckless. Then on February 2nd, 2026, Musk made a move that rewrote the record books. SpaceX announced it had acquired XAI. The deal valued SpaceX at $1 trillion and XAI at $250 billion for a combined value of $1.25 trillion. According to data compiled by LSG and reported by Reuters, it was the largest merger or acquisition in history. It broke a record that had stood for more than 25 years set when Vodafone bought Germany's manisman for $23 billion in 2000. Investors in XAI received 0.1433 shares of SpaceX for every share of XAI they owned. Musk described the deal with his usual flare. He said it marked not just the next chapter, but the next book in the company's mission, scaling to make what he called a sentient sun to understand the universe. Underneath the poetry, the business logic was simpler. According to CNBC, Musk said the main reason for the merger was to build orbital data centers, computers, and space. Another reason was money. XAI needed huge amounts of capital, and SpaceX with Starlink throwing off cash could provide it. In March, SpaceX took out a $20 billion bridge loan at a much cheaper interest rate and used it to pay off XAI's debt, moving that debt onto SpaceX's own balance sheet. Critics raised concerns right away. Reuters noted that a deal like this could draw scrutiny from regulators and investors over governance valuation and conflicts of interest because Musk leads several companies at once, and SpaceX holds billions of dollars in contracts with NASA, the Department of Defense, and intelligence agencies. But the merger went through, and it set up the biggest stock market debut the world has ever seen. On June 12th, 2226, SpaceX began trading on the NASDAQ under the ticker SPCX. The IPO raised $75 billion at $135 a share. That was more than double the previous record set by Saudi Aramco in 2019 on its first day. The stock jumped 19% and closed at $160.95, valuing the company at about $2.1 trillion. According to Reuters, the gain made Elon Musk the world's first trillionaire. retail investors. Regular people received about 20% of the shares, far more than in a typical IPO. Not everyone cheered. On the day of the debut, Reuters reported that SpaceX was trading at roughly 112 times its revenue, far above other mega cap stocks. Morning Star analysts said the company was more fairly valued at around $780 billion. And the research firm CFRA started coverage with a sell rating. One investment chief said plainly that this was not a name you buy based on fundamentals. So the stage was set. A company valued at about $2 trillion with a giant AI business attached that was still losing money. The question everyone was asking was simple. Would the AI bet ever pay off? In the first earnings report after the IPO, the answer looked shaky. For the second quarter, which ended in June, SpaceX's AI segment brought in $2.56 billion in revenue, but it lost $1.26 billion from operations. The company as a whole posted a net loss of $541 million. Then came October 20th. SpaceX's third quarter results covering July through September were nothing like the quarter before. Total revenue was 14.58 billion, up 161% from 5.58 billion a year earlier. Analysts had expected around 12.9 billion. The company posted a net income of $4.4 billion compared with a net loss of about $1.1 billion a year earlier. Wall Street had expected profit of around 1.6 billion. Operating income came in at $2.79 billion. The company's first operating profit as a public company, the AI segment, which SpaceX now calls super intelligence or SI, brought in $7.56 billion. That's up 195% from the previous quarter and 830% from a year ago. Analysts had expected about 6.47 billion. For the first time, the SI segment turned an operating profit $4301 million. And that's after $1.2 billion in stockbased compensation tied to an acquisition I'll explain in a moment. Starlink was strong, too. And that's what makes the shift so striking. The connectivity segment, mostly Starlink, brought in $5.25 25 billion, up 69% from a year earlier. Its operating income nearly doubled to 2.37 billion. Starlink reached 13.6 million consumer subscribers, adding 1.6 million in a single quarter. Even with all that growth, the AI business passed it. The space segment, the rockets that made SpaceX famous brought in $1.77 billion in the quarter, up 6% from a year ago, and posted a small operating loss of $7 million. That's not because launches stopped. SpaceX flew 35 times during the quarter, including Starship Flight 14, its first orbital Starship flight, which carried 26 Starlink V3 satellites. The space segment is losing money because SpaceX is pouring huge sums into Starship development. So, here's the reality in plane numbers. In the third quarter of 2026, a little more than half of SpaceX's revenue came from the business most people don't think of when they hear the name SpaceX. That brings us to the surprise I mentioned at the start. Where does that $7.6 billion actually come from? Most people assume it's Grock. It's the most visible AI product Musk owns. It's built into X and it's in the headlines all the time. But according to Space X's own earnings release, only $538 million of the SI segment's revenue came from advertising, which is mostly the Xplatform. The other $7.02 billion came from a category SpaceX calls SI solutions and infrastructure. And according to Reuters, most of that comes from renting out data center capacity to other AI developers rather than from SpaceX's own models. Reuters named those customers Anthropic Alphabets, Google, and Reflection AI. So SpaceX is making billions of dollars by renting its computers to the companies it competes with. The biggest of those deals is with Anthropic, the company behind the clawed AI models. In May, Anthropic announced it would use all of the compute capacity at SpaceX's Colossus 1 data center in Memphis, Tennessee. According to SpaceX's AI unit, Colossus 1 has more than 220,000 Nvidia GPUs, including H100, H200, and next generation GB200 chips. Anthropic soon expanded into Colossus 2 as well. Then the IPO filing revealed the price, and it was huge. According to Axios, Anthropic is paying SpaceX $1.25 billion per month through May 2029. That works out to roughly $15 billion a year from a single customer. There's an important catch. According to the filing, either company can exit the deal with 90 days notice. Musk himself later posted that SpaceX had only agreed to a 180day lease with a mutual 90-day cancellation notice after that. He said the short term was SpaceX's request, not Anthropics, and that if compute gets super tight, SpaceX might need it back at some point. So, a large share of this revenue rests on contracts that on paper are not locked in for years. Reuters summed up the shift well. For now, it said SpaceX's main AI business looks closer to a cloud providers than an AI labs. In other words, SpaceX is becoming a landlord for the AI boom, while companies like anthropic Google and OpenAI race to build smarter models. They all need the same scarce resource, enormous amounts of computing power, plus the electricity to run it. SpaceX built that capacity fast, and now it's renting it out. The SI segment's growth didn't only come from rented computers. During the quarter on August 14th, SpaceX completed its acquisition of Cursor, the popular AI coding tool, at an implied equity value of $60 billion. From the day the deal closed through the end of September, Curser added $68 million of revenue to SpaceX's results. SpaceX also launched what it calls Grockbot, an always an AI agent, and said Grock adoption was growing among businesses and consumers. Even so, the rental business is the main driver. Now, let's look at the other side of the ledger, because this is where the story gets dramatic. To earn $7.6 billion from AI in one quarter, SpaceX spent $16.4 billion on AI capital expenditures in that same quarter across the whole company capital spending hit $9.6 billion, nearly triple the 6.9 billion it spent a year earlier. And this is the number that made analysts sit up. SpaceX's chief financial officer, Brett Johnson, told investors the company expects to spend 35 to40 billion in the fourth quarter alone. Think about the scale of that. $40 billion is more than the entire annual economy of many countries. And SpaceX may spend it in three months, mostly on chips, power cooling, and buildings for AI. Where's it all going? SpaceX measures its AI capacity in what it calls nameplate compute. The total power its computing hardware is rated to draw. As of September 30th, SpaceX reported 1.7 gawatts of name plate compute. 3 months earlier, it was 1.4 gawatt. A year earlier, just 0.6. For a sense of scale, 1 gawatt is roughly the output of a large nuclear power reactor. SpaceX says the growth came from building out Colossus SKS. Some Wall Street analysts expect even more, according to Fortune, citing S&P Global's visible alpha research. Consensus estimates have SpaceX's total capital spending rising from about $70 billion this year to roughly 197 billion in fiscal 2028. The same research projects SpaceX's debt could rise from about $50 billion in 2025 to around 350 billion in 2029. Those are analyst projections, not company guidance, but they show what's at stake. This is one of the most capital-hungry bets any company has ever made. So why are investors willing to fund it? because of the numbers Musk is promising next on the earnings call. Johnson reaffirmed a target of 100 billion in annualized revenue by the end of this year. And Musk in his usual style went much further. He said SpaceX's buildout pace implies annual recurring revenue of 500 billion to1 trillion by the end of 2027. Let's check that math because this channel doesn't just repeat claims. Fortune did the calculation. At its third quarter pace, SpaceX is bringing in roughly $58 billion a year. To hit a $100 billion annualized run rate, it would need about $8.3 billion in revenue in December alone, roughly 70% more than its average month last quarter. Musk's range for the end of 2027 would require somewhere between 42 and $83 billion a month. Today, the entire company brings in a bit under five billion a month. Musk has a long history of setting timelines that slip. He also has a history of building things people said couldn't be done. You've seen both on this channel. So treat those targets as goals, not facts, and watch the actual quarterly numbers. There's one more piece to the strategy, and it's the one that connects AI back to rockets. It's called Star Mind. Star Mind is SpaceX's plan for orbital data centers, computers that live in space. The idea is that in orbit, solar power is nearly constant and abundant. And the company argues that space offers what it calls near limitless sustainable power with less impact on Earth. SpaceX has filed with the Federal Communications Commission for a constellation of up to 1 million of these AI satellites. About 100 times larger than the current Starlink network. The FCC is still reviewing the application. According to PCMAG, each satellite could have a mass of up to 4,000 kg with solar panels spanning up to 246 ft. SpaceX's senior vice president, Michael Nichols, said the company plans to launch the first Starmine satellites next year, and that Star Mind will be the largest consumer of Starship's launch capacity. Here's where the whole empire clicks together. Starship launches the satellites. Starlink's laser links carry the data. The AI business uses the compute and customers like Enthropic have already expressed interest in partnering on multiple gigawatts of orbital AI compute capacity. The plan has serious critics. Astronomers warn it could flood the night sky with light pollution. The University of Virginia astronomy professor Samantha Lawler, who co-authored research on orbital data centers, told PCMAG that her team simulations show more satellites than stars visible in the sky for most of the world for most of the night and for most of the year. Others have raised concerns about space debris and collisions, especially since the satellites are designed to last only about 5 years. Even Nichols described SpaceX's own video of the constellation as a very notional concept that's probably wrong and will change a lot. Starmine doesn't exist yet. It's a plan. So now you have the full picture of SpaceX in late 2026. A rocket company that bought an AI company, went public at a value of about $2 trillion, and is now earning more from AI than from anything else while spending at a pace almost no company in history has matched. Which brings us to Tesla. Why would a company like this want an electric car maker? Start with what Musk actually said on the call. In his words, I do think there's a lot of merit to combining the companies. As we are pretty much in the super intelligence singularity and there's a lot to be gained from cooperating on developing super intelligence software chip design the terapab obviously robotics mass orbit with the star mine constellation connectivity each item in that list is a real overlap so let's go through them the first is the terraab this is a giant chip manufacturing complex that SpaceX and Tesla are building together in Grimes County Texas according to Reuters the companies will initially invest $16.8 8 billion. An earlier May filing showed a proposed initial investment of $55 billion, rising to as much as 119 billion if all phases are completed. The plant is planned at 100 million square feet and is meant to make package and test advanced logic and memory chips under one roof for Tesla's Optimus robots and cyber cabs and for SpaceX's space-based data centers. It's expected to employ at least 3,000 people and Intel as a partner. on the call. Musk said Terafab should produce its first test chip around May or June of next year and aims to reach volume production by the third quarter of 2028. The second overlap is robotics and physical AI. Tesla is pivoting hard away from being just a car company. In January, Musk said Tesla would phase out production of its Model S and Model X. Tesla is now focused on autonomous products, self-driving robo taxis, the two seat cyber cab, and the Optimus humanoid robot. All of that depends on AI. Barklay's analyst Dan Levy described the logic this way. Tesla's physical AI like robo taxi and Optimus would be paired with SpaceX's virtual AI like Grock, all supplied by the shared terapab chip plant. The third overlap is money that's already moving between them. Tesla invested $2 billion in XAI in January 2026 before XAI became part of SpaceX. According to Fortune, Tesla's latest proxy statement confirms that investment, which now effectively means Tesla owns a piece of Space X. The boards overlap, too. Fortune reported that venture capitalist Ira Aaron Price chairs SpaceX's compensation and nominating committee as well as the two separate equivalent committees at Tesla. Musk has also pointed to smaller links like SpaceX buying Tesla Cybert trucks. at the all-in summit in September. He said, "With all this collaboration on so many levels, who can imagine what action one might take when there's so much close collaboration in so many areas?" And then there's the reason that's rarely said out loud, but that every analyst is discussing control. According to Barclays, Musk owns around 20% of Tesla, but about 48% of SpaceX and SpaceX's super voting class B shares give him around 82% of the voting power there. For years, Musk has said he wants about 25% voting control of Tesla so that in his view, he can't be pushed out as he builds robots and AI. Barkclay has ran the numbers. If SpaceX bought Tesla in an all stock deal at current values, Musk would own about 37% of the combined company and control about 56% of the vote. That's far above the 25% he has long said he wants. Reuters reported the same conclusion from analysts. A merger would give Musk control of the combined entity. So, which company would buy which? According to a person familiar with the matter who spoke to Reuters, the talk started in recent weeks and could result in SpaceX acquiring Tesla. Tesla has been working with an investment bank to advise on the discussions. Barclays also expects SpaceX to be the buyer in an all stock deal. At recent values, Bararkclays estimated each Tesla share would be worth about 0.7 SpaceX shares before any takeover premium. That premium is where the numbers get eyewatering. At closing prices on October 20th, Fortune calculated that a combined Tesla and SpaceX would be worth about $3.5 trillion with Musk owning about a third. But Dan Ies of Yorkville Ives and Company told CBS News that Tesla shareholders would likely demand a premium to approve a deal likely in the $500 to $550 per share range in his view. That would push the total value of the deal toward $5 trillion. That would make it the largest corporate deal in history and it would put the combined company in the same league as Nvidia and Apple. Ives puts the odds of a merger happening by the end of 2027 at 80 to 90%. The market reacted fast the morning after the call. According to CBS News, SpaceX shares jumped nearly 10% to about $162 and Tesla rose almost 4% to about 388. But this is the part most headlines skipped and it's the most important part of the whole story. Musk himself warned against assuming the deal is done. In his words, I think some people out there assume that a SpaceX and Tesla merger is a certainty, and I think that would be overstating it. He said the special committees had not agreed on any terms of transaction and may never do so. He also said he can't take part in their discussions because he's a conflicted party since he controls both sides. He called the whole situation an odd kabuki dance where two groups of people are talking and I can't be in the conversation. Then he described the hurdle that could decide everything. On the Tesla side, he said, "We will need a 50% vote of all shares. Not just those who vote, but of all shareholders. And that's quite a high hurdle to try to achieve." That detail matters more than it sounds. In most shareholder votes, the result depends on the people who actually show up and vote. Here, according to Musk, a merger would need support from a majority of all of Tesla's outstanding shares. Every share held by someone who doesn't bother to vote effectively counts. as a no. Tesla has millions of retail shareholders and getting enough of them to participate is genuinely difficult. There's also a clock. Barkclay has pointed out that Tesla must hold its annual shareholder meeting by December 6th under SEC rules and suggested that meeting could be the natural place for a vote. But Fortune reported that the Tesla proxy filed on October 19th does not include a merger proposal. Barkclays warned that if an announcement slipped later into October or early November, approval at that meeting would become increasingly difficult. In that case, a deal might come in the next 6 to 12 months instead. So, what does Tesla itself look like right now? The company Musk might combine with SpaceX. One day after the SpaceX call on October 21st, Tesla reported its own third quarter results, and they show why the AI story matters so much to Tesla investors. Tesla's revenue was $28.05 05 billion, roughly flat from a year earlier and slightly above estimates. Adjusted earnings were 48 cents a share, beating the 43 cents analysts expected, but the core car business shrank. Automotive revenue fell 3% to 20.5 to6 billion, and energy revenue fell 16%. Net income dropped 29% to $787 million as spending on Optimus robots and other AI projects surged. Capital spending more than doubled to $3.73 billion. Vehicle deliveries in the quarter were 486,532, about 2% below a year earlier. As Chinese automakers like BYD and Shyomi keep gaining ground, there were bright spots, and they're all about AI. Tesla said active subscriptions to its full self-driving supervised system rose 67%. from a year earlier to 1.74 million. According to Texas's Department of Motor Vehicles, as reported by CNBC, Tesla's driverless fleet includes 362 Model Y robo taxis and 449 cyber cabs authorized for driverless operation in Texas. Musk said the robo taxi business won't move the needle this quarter, but should become noticeable in the second quarter of next year and probably a big deal by the third or fourth quarter of 2027. On Optimus, he admitted there's still quite a lot of work to do before the robot is truly useful for open-ended tasks and said he expects the main problems to be solved around the end of next year. And on the merger, Musk said almost nothing on the Tesla call. In his words, I think I said everything that should be said yesterday. Put the two companies side by side and you can see why the idea is so tempting and so controversial. SpaceX's revenue is growing at tripledigit rates. It's newly profitable and it's building AI infrastructure at enormous scale. Tesla is a car company whose car business is shrinking, betting its future on robots, robo taxes, and AI chips. Both need vast numbers of chips and they're building the same chip factory. One man runs both. So, is this a brilliant integration of a single AI and robotics empire or a risky concentration of power and money that shareholders should be wary of? Let's lay out both sides fairly because that's what this channel does. The supporters case goes like this. AI is moving toward the physical world, robots and self-driving vehicles. And the companies that control the full stack will have an advantage. That stack runs from chips to data centers to models to the machines that use them. Musk already controls more pieces of it than almost anyone. Merging would let SpaceX and Tesla share the terapab share. AI models share engineers and avoid the conflicts of interest that come with constant deals between two Musk companies. Investors have generally rewarded Musk for thinking big. Dan Ies called it the right move for Musk for a number of strategic AI and datadriven reasons. The skeptics's case is just as serious. First, the valuations. SpaceX traded at more than 100 times its revenue at the IPO. And Morning Star argued it was worth far less than the market says. Second, the spending. $40 billion of capital spending in one quarter. And analyst projections of hundreds of billions in debt means SpaceX's AI business has to keep growing at extraordinary speed to justify it. Third customer concentration. A big share of SI revenue comes from renting compute to a small number of AI companies under contracts that can be ended on 90 days notice. If those companies build their own capacity or if AI demand cools that revenue could shrink fast. Fourth, governance. Musk controls both companies and a merger would increase his control further. Critics argue that reduces the power of other shareholders to check his decisions. And fifth, Musk himself is a lightning rod. In the same week as the earnings report, the Wall Street Journal reported that Musk ordered a slowdown of a military-grade version of Starlink, a report CNN said it could not immediately verify. And Ukraine accused SpaceX of hindering its defense by limiting Starlink service. Musk has said that limits on the military Star Shield service are decided by the US government. Whatever the truth, controversies like this show how much risk is concentrated in one person. And there's a bigger question underneath all of this. The one that should matter to anyone who follows space. If SpaceX becomes mainly an AI company, what happens to the space mission? Right now, the numbers say the space business is the smallest of SpaceX's three segments and the only one that lost money last quarter. But it may be the most important for the future because Starship is the key to everything else. It's the rocket that's supposed to launch the bigger Starlink satellites, the Star Mine data centers, and NASA's Aremis astronauts to the surface of the moon. The company also announced plans for a new spaceport Starbase Louisiana designed to support thousands of Starship flights a year with the first launch targeted for 2029. The AI money is now paying for the rockets and the rockets are supposed to carry the AI into orbit. Each part depends on the others. That's either a beautifully designed flywheel or a tower of dependencies depending on who you ask. So where does this leave us as of right now? SpaceX's super intelligence division brought in 7.56 billion. Last quarter up 830% in a year. Most of that came from renting computing power to other AI companies, including a competitor paying about $1.25 billion a month. It passed Starlink for the first time and turned a profit for the first time. SpaceX as a whole earned $4.4 billion after back-to-back losing quarters. At the same time, it's spending money at a pace almost no company has ever matched with up to $40 billion of capital spending expected in a single quarter. And Elon Musk has confirmed for the first time that SpaceX and Tesla have formed special committees to discuss a merger he believes is a good idea. That deal could be worth $3.5 to 5 trillion and could give him majority voting control of the combined company. But no terms have been agreed. Musk himself says it's not a certainty and on the Tesla side, it would need a majority of all outstanding shares. a hurdle he calls quite high. Here's what to watch next. First, whether a formal merger proposal appears before Tesla's annual meeting, which must be held by December 6th. Second, SpaceX's fourth quarter results, which will show whether it actually hit its 100 billion annualized revenue target, and how much of that massive capital spending turned into new revenue. Third, whether the big compute customers, especially anthropic, stay on or start pulling capacity back. And fourth, the first terapab test chip next spring and the first Starmine satellites next year, which would move the story from plans to hardware. Four years ago, SpaceX was a rocket company with a satellite internet business. Today, by revenue, it's an AI company that also happens to launch most of the world's rockets. And if Musk gets his way, it may soon also build the world's robo taxes and humanoid robots. Whether that becomes the most powerful technology company ever built or one of the most expensive bets in corporate history may be decided by a vote of shareholders who have never been asked a bigger question. If this breakdown helped you understand what's really happening inside SpaceX, hit like subscribe and turn on notifications. Because the moment a formal merger proposal is filed, I'll break down exactly what's in it, what's confirmed, and what isn't. Tell me in the comments, do you think SpaceX and Tesla should merge? And would you vote yes if you own Tesla shares? I read the answers. And if you want to see the rocket that all of this ultimately depends on, watch my breakdown of Starship's first orbital flight next. Because without it, none of the AI in space plan gets off the